Search results

[Bug]: forward_capture AttributeError ('NoneType' has no attribute 'extend') during MoE calibration when a kwarg is None-then-Tensor

Status: Open.

#1950 In intel/auto-round;

· pasta-paul opened on Jun 24, 2026

[Bug]: autoscheme gguf format is less accurate than fake.

bugSomething isn't workingSomething isn't working

Status: Open.

#1948 In intel/auto-round;

· n1ck-guo opened on Jun 24, 2026

[Bug]: Segmentation fault occurred when use auto round quant GLM-4.7-FP8

bugSomething isn't workingSomething isn't working

Status: Open.

#1946 In intel/auto-round;

· fxu11x opened on Jun 23, 2026

[Feature]: Consolidate three AutoRound classes

enhancementNew feature or requestNew feature or request

Status: Open.

#1944 In intel/auto-round;

· wenhuach21 opened on Jun 23, 2026

[Feature]: Introduce built-in ar, ar-light, and ar-best recipes for each quantization scheme

enhancementNew feature or requestNew feature or request

Status: Open.

#1940 In intel/auto-round;

· wenhuach21 opened on Jun 22, 2026

·  

0.15.0

[Bug]: Accuracy issue observed on G2E with quantized minimax m2.7

bugSomething isn't workingSomething isn't working

high priority

Status: Open.

#1934 In intel/auto-round;

· fxu11x opened on Jun 17, 2026

·  

0.14.0

[Bug]: DS-R1 mxfp4 model vllm eval failed

bugSomething isn't workingSomething isn't working

Status: Open.

#1932 In intel/auto-round;

· chensuyue opened on Jun 17, 2026

[Bug]: gemma-4-E4B-it gguf acc=0

bugSomething isn't workingSomething isn't working

Status: Open.

#1926 In intel/auto-round;

· XuehaoSun opened on Jun 17, 2026

·  

0.14.0

Production report: Qwen3.6-35B-A3B int4-mixed replaced a 4.75-bit build on DGX Spark (GB10), +12.7% decode, no quality loss, vision intact

Status: Open.

#1919 In intel/auto-round;

· cipherfoxie opened on Jun 12, 2026

[Bug]: Qwen3-32B failed with mxfp4 quant on BMG

bugSomething isn't workingSomething isn't working

high priority

Status: Open.

#1918 In intel/auto-round;

· chensuyue opened on Jun 12, 2026

·  

0.14.0

[Bug]: autocheme bits do not correctly be set in gguf

bugSomething isn't workingSomething isn't working

Status: Open.

#1917 In intel/auto-round;

· wenhuach21 opened on Jun 12, 2026

[Bug]: Use auto round quant minimax m2.7, the moe gate weight will be quanted as default.

bugSomething isn't workingSomething isn't working

high priority

Status: Open.

#1915 In intel/auto-round;

· a32543254 opened on Jun 11, 2026

·  

0.14.0